Type of discrepancy between glycosylated hemoglobin and fasting plasma glucose is associated with in-hospital outcomes in patients with acute coronary syndrome and diabetes: �ndings from the Improving Care for Cardiovascular Disease in China - Acute Coronary Syndrome (CCC-ACS) Project

; 95


Introduction
Cardiovascular disease is the leading cause of both death and premature death in China, being the cause of 40% of deaths in the Chinese population [1].Acute coronary syndrome (ACS) is an acute manifestation of cardiovascular disease with a high risk of mortality, which can lead to critical conditions such as cardiogenic shock and cardiac arrest.Patients with ACS and diabetes usually have worse clinical outcomes than those with normal blood glucose [2][3][4][5][6][7][8], regardless of in-hospital or long-term outcomes.
Indicators commonly used for evaluating blood glucose include intravenous blood glucose, glycosylated hemoglobin (HbA 1c ), and glycosylated serum albumin.Intravenous blood glucose was rstly used in the diagnosis of diabetes, including fasting plasma glucose (FPG), oral glucose tolerance test and random blood glucose.In 2013, the American Diabetes Association approved the use of HbA 1c to diagnose diabetes [9].In addition to diagnostic value, FPG and HbA 1c are also associated with clinical outcomes.
However, we noticed that some patients have a discrepancy between HbA 1c and FPG in clinical practice.This condition may be related to factors such as acute stress, renal dysfunction, anemia that may affect FPG and HbA 1c .It may be an increased FPG with normal HbA 1c , or an increased HbA 1c with normal FPG.
We are confused that which one indicates worse in-hospital outcomes in this case, no matter what causes the discrepancy.There are few studies focusing on this issue.
The Improving Care for Cardiovascular Disease in China -Acute Coronary Syndrome (CCC-ACS) project is a national, hospital-based quality improvement project with an ongoing database, aiming to increase adherence to ACS guidelines in China and to improve patient outcomes.We conducted this study based on CCC-ACS project to investigate the types of discrepancy between HbA 1c and FPG and its relationship to in-hospital outcomes.

Research design
The datasets used and analyzed during the current study are available from the principal investigator of CCC-ACS on reasonable request.Details of the design and methodology of the CCC-ACS project have been published [18].In brie y, it is a national, hospital-based quality improvement project with an ongoing database, aiming to increase adherence to ACS guidelines in China and to improve patient outcomes.It was launched in 2014 as a collaborative initiative of the American Heart Association and Chinese Society of Cardiology.240 hospitals were recruited representing the diversity of ACS care in hospitals in China, including 150 tertiary hospitals in phase and phase , 82 secondary hospitals and 8 tertiary hospitals in phase (from July 2017) and phase (from November 2018).Clinical data were collected via a webbased data collection platform (Oracle Clinical Remote Data Capture, Oracle Corporation).Trained data abstractors entered the data elements abstracted from medical charts.Eligible patients were consecutively reported to the CCC-ACS database for each month before the middle of the following month.Around 5% of reported cases were randomly selected and compared with the original medical records.An audit by a third party was performed to ensure that cases were reported consecutively rather than selectively.

Research population
A total of 104,516 inpatients with ACS, identi ed using their principal diagnosis at discharge, were enrolled from 240 hospitals across China from November 2014 to December 2019.Patients with diabetes and complete HbA 1c and FPG value at admission were included in this study.Only patients from July 2017 to December 2019 were included in this study because that the FPG value at admission was not included in the database before July 2017.Patients were divided into consistent group and discrepancy group based on the HbA 1c and FPG value at admission.Consistent group included patients with HbA 1c < 6.5% and FPG < 7.0 mmol/L, or patients with HbA 1c ≥ 6.5% and FPG ≥ 7.0 mmol/L.Discrepancy group included patients with HbA 1c ≥ 6.5% and FPG < 7.0 mmol/L, or patients with HbA1c < 6.5% and FPG ≥ 7.0 mmol/L.Discrepancy group was further divided into increased HbA 1c but normal FPG group (HbA 1c ≥ 6.5% and FPG < 7.0 mmol/L) and increased FPG but normal HbA 1c group (HbA 1c < 6.5% and FPG ≥ 7.0 mmol/L).Institutional review board approval was granted for the aggregate data set for research and quality improvement by the Ethics Committee of Beijing Anzhen Hospital, Capital Medical University.No informed consent was required.

De nition of variables
diabetes was de ned as having a history of diabetes, or receiving glucose-lowering agents before hospitalization, or diabetes listed in the medical records as the secondary discharge diagnosis, or HbA 1c ≥ 6.5% at admission.Hypertension was de ned as having a history of hypertension, or receiving antihypertensive medication, or systolic blood pressure ≥ 140 mmHg or diastolic blood pressure ≥ 90 mmHg at admission.The ACS classi cation was based on the primary diagnosis of discharge in the medical record.Non-ST-segment elevation ACS (NSTE-ACS) was de ned as non-ST-segment elevation myocardial infarction (STEMI) or unstable angina.All the laboratory testing values were the values tested rst time after admission.The estimated glomerular ltration rate (eGFR) was calculated according to the equation developed by the Chronic Kidney Disease Epidemiology Collaboration [19].The medication was prescribed after admission.

In-hospital outcomes
The outcomes of this study included major adverse cardiovascular and cerebrovascular event (MACCE), heart failure, composite of cardiovascular death and heart failure, composite of MACCE and heart failure, and death from any cause.MACCE was de ned as cardiovascular death, cardiac arrest, cardiogenic shock, recurrent myocardial infarction, stent thrombosis, and stroke.

Statistical analysis
Continuous variables were presented as mean and standard deviation (SD) or median and interquartile range (IQR) when distribution and variance met the conditions.The categorical variables were presented as a percentage.

Characteristics of patients in discrepancy group
A total of 7,762 patients were included in this study (Additional le 1: Fig. S1).The mean age was 64.The proportion of increased FPG but normal HbA 1c group was higher in patients with age over 65 years, hemoglobin less than 120 g/L, eGFR less than 60 ml min − 1 (1.73 m) −2 , Killip class III or IV, STEMI, and were treated with oral glucose-lowering drugs or insulin (Fig. 1, Additional le 1: Table S2).

Types of discrepancy and in-hospital outcomes
The comparison of in-hospital outcomes between increased HbA 1c but normal FPG group and increased FPG but normal HbA 1c group were shown in Fig. 2. The rates of all the in-hospital outcomes were higher in increased FPG but normal HbA 1c group than in increased HbA 1c but normal FPG group.Logistic regression model was performed in order to explore the relationship between the types of discrepancy and in-hospital outcomes, except for the death from any cause because of the small event number.In univariate logistic regression analysis, a signi cantly higher risk of all the in-hospital outcomes was observed in patients of increased FPG but normal HbA 1c group ( Subgroup analysis was performed based on age, gender, medical history, Killip class, hemoglobin, eGFR, type of ACS, and glucose-lowering drug use before hospitalization.Higher risk of all the in-hospital outcomes was observed in patients of increased FPG but normal HbA 1c group, which was consistent in all subgroups (all P for interaction > 0.05), except for the eGFR subgroup for MACCE (Fig. 3).which shown that increased FPG but normal HbA 1c group increased the risk of MACCE more clearly in patients with eGFR ≥ 60 ml min − 1 (1.73 m) −2 .

Discussion
In this study, we investigated the types of discrepancy between HbA 1c and FPG in patients with ACS and diabetes.We found that nearly one third of patients had a discrepancy between HbA 1c and FPG.Of these patients with discrepancy, the patients with increased FPG had higher risk of in-hospital cardiovascular adverse outcomes than those with increased HbA 1c .
Discrepancy between HbA 1c and FPG was reported by some studies.A study about the risk of hypertension in patients with prediabetes demonstrated the discrepancy HbA 1c and FPG [20].A study using data from residents of Yunnan Province, China, showed that the discrepancy between HbA 1c and FPG occurred in about 30% of participants [21].In our study, the discrepancy between HbA 1c and FPG can be also found in patients with ACS and diabetes.We found that discrepancy group, which included 77.5% of patients in increased HbA 1c but normal FPG group and 22.5% of patients in increased FPG but normal HbA 1c group, accounted for 29% of total study population.Patients often experience hyperglycemia in the acute phase of many diseases, such as ACS, which is called stress hyperglycemia.
HbA 1c re ects average glycemia over approximately 3 months, so an increase in HbA 1c usually indicates chronic hyperglycemia.We found that patients in increased FPG but normal HbA 1c group were more likely to have lower eGFR and be treated with glucose-lowering agents.Higher proportion of glucose-lowering agents use may be related to the well-controlled blood glucose and lower HbA 1c .Furthermore, changes in the metabolism of glucose-lowering drugs, insulin clearance, and the uremic environment in patients with renal function insu cient may also lower HbA 1c value [22].From our study, not only can the discrepancy between HbA 1c and FPG be found in patients with chronic kidney disease (CKD), but also the proportion of increased FPG group was signi cantly higher than that of increased HbA 1c group.
There is a strong association between cardiovascular disease, diabetes and CKD.People with diabetes and CKD have a greatly increased risk of all-cause mortality, cardiovascular mortality, and kidney failure [23,24].Furthermore, we analyzed the relationship between the types of discrepancy and in-hospital outcomes.We have known that HbA 1c and FPG were both closely related to the in-hospital outcomes.
Most previous studies have shown that increased HbA 1c or FPG was signi cantly associated with poor inhospital outcomes in patients with ACS and diabetes.An observational study included 250 patients with ACS, which found that coronary atherosclerosis was more advanced in patients with HbA 1c ≥ 5.7% than those with HbA 1c < 5.7% [17].Goyal, et al [25] conducted a post hoc analysis including two randomized controlled trials of acute myocardial infarction with ST-segment elevation, involved 30,536 subjects with diabetes history, which showed that patients with in-hospital glucose ≥ 144 mg/dL had a very high risk of death.However, in clinical practice, some conditions such as acute stress, renal dysfunction, and anemia can cause uncertainty in the measured values of FPG and HbA 1c , such as discrepancy between FPG and HbA 1c .Until now, it is not so clear about the association of in-hospital outcomes with the discrepancy between HbA 1c and FPG in patients with ACS and diabetes.There are few studies focusing on this issue.
From our study, we can draw a conclusion that patients in increased FPG group, who were more likely to have higher heart rate, poorer heart function, higher incidence of STEMI as well as hypertension, had higher risk of in-hospital cardiovascular adverse outcomes than those with increased HbA 1c .Stress hyperglycemia, which is a re ection of high free fatty acids, insulin resistance, and steroid hormones, affects the course of the disease in the worst way [26].From other study, we have learned that the level of stress hyperglycemia often correlates with the severity of disease and predict mortality [27].In our study we also found that the patients with severe clinical condition, such as the higher heart rate and the poorer heart function, were more likely to have an increase in FPG.As a result, stress hyperglycemia may have greater adverse effect on patients with ACS and diabetes than chronic hyperglycemia.
The ndings of this study may have some important implications for clinical practice.The HbA 1c test is the major tool for assessing glycemic control and has strong predictive value for diabetes complications [28].Chronic hyperglycemia is an important risk factor for cardiovascular disease and mortality [24], although the variability of HbA 1c in patients with renal insu ciency should be concerned.However, in patients with ACS and diabetes, increased FPG may be associated with the higher risk of adverse inhospital outcomes, even though the HbA 1c is well controlled.These patients, especially including those with renal insu ciency, should be given more attention and closer monitoring in clinical practice.
The major strength of our study is that it is based on a nationally representative registry and aimed at investigating the discrepancy between HbA 1c and FPG and the in uence on the in-hospital outcomes of patients with ACS and diabetes, which was rarely reported till now.Our study also has certain limitations.Firstly, all-cause mortality was not included in the logistic regression analysis because of very limited events.Secondly, we cannot collect all information affected glucose metabolism from this real-world research for ACS patients based on medical records, thus contributing to some residual confounding from unmeasured confounders.Lastly, fasting statue, blood sample collection and testing methods were di cult to unify, as this was a real-world multicenter study.

Conclusion
In summary, our study showed that the discrepancy between HbA 1c and FPG accounts for nearly 30% in patients with ACS and diabetes.Patients with an increased level of FPG had a higher risk of in-hospital cardiovascular adverse outcomes than those with an increased level of HbA 1c .This result may indicate that when HbA 1c and FPG are inconsistent in patients with ACS and diabetes, the increased FPG that may be caused by stress hyperglycemia may have greater adverse effect than those with increased HbA 1c that may be caused by chronic hyperglycemia.These high-risk patients, especially including those with renal insu ciency, should be given more attention and closer monitoring in clinical practice.
In-hospital outcomes in patients with discrepancy between HbA1c and FPG.HbA1c, glycosylated hemoglobin; FPG, fasting plasma glucose; MACCE, major adverse cardiovascular and cerebrovascular event.

Figure 3 Subgroup
Figure 3 , gender, blood pressure, heart rate, eGFR, hemoglobin, heart function, type of ACS, medical history, glucose-lowering drug use, in-hospital management, and referral status.The heterogeneity of effects on the in-hospital outcomes across subgroups was estimated using random effects meta-analysis.For data with missing value less than 15% (Additional le 1: TableS1), sequential regression multiple imputation method implemented by IVEware software version 0.2 (Survey Research Center, University of Michigan, Ann Arbor, MI, USA) was used to impute the missing value.All P values were 2-tailed and a P < 0.05 was considered statistically signi cant.Statistical analyses were performed using SPSS 23.0 (SPSS Inc., Chicago, IL) and Stata/IC 15.1.
The comparison between groups of continuous variables was performed by unpaired ttest or Mann-Whitney U test (Kruskal-Wallis), and the chi-square test was used to compare the categorical variables.Multivariate logistic regression model was used to determine the association between the types of discrepancy and in-hospital outcomes by controlling potentially confounders.Candidate adjustment factors included age

Table 1
Characteristics of patients with discrepancy between HbA 1c and FPG.
4 (± 11.6) years, males accounted for 68.8%.The mean hemoglobin was 135.1 (± 21.3) g/L and mean eGFR was 81.7 (± 25.4) ml min − 1 (1.73 m) −2 .53.3% of patients were treated with at least one class of oral glucose-lowering drug or insulin.The patients in consistent group and discrepancy group were 5,490 (70.7%) and 2,272 (29.3%) respectively.Patients in discrepancy group were more likely to have lower eGFR (Additional le 1: Fig. S2).In discrepancy group, increased HbA 1c but normal FPG group accounted for 77.5% (1,761/2,272), and increased FPG but normal HbA 1c group accounted for 22.5% (511/2,272).The baseline characteristics for patients in increased HbA 1c but normal FPG group and increased FPG but normal HbA 1c group were shown in Table 1.Patients in increased FPG but normal HbA 1c group were more likely to have lower eGFR, higher heart rate, poorer heart function, have STEMI and hypertension, and be treated with glucose-lowering agents.1c , glycosylated hemoglobin; FPG, fasting plasma glucose; SD, standard deviation; IQR, interquartile range; CHD, coronary heart disease; ACS, acute coronary syndrome; STEMI, ST-segment elevation myocardial infarction; NSTE-ACS, non-ST-segment elevation acute coronary syndrome; eGFR, estimated glomerular ltration rate; HDL, high density lipoprotein; LDL, low density lipoprotein; ACE, angiotensin-converting enzyme.

Table 2
Logistic regression analysis for in-hospital outcomes in increased FPG group compared with increased HbA 1c group.